Any fixed parity of size at least logarithmic in the dimension requires exponentially many perturbed-gradient steps before the expected correlation loss moves away from its trivial value.
Further and stronger analogy between sampling and optimization: Langevin monte carlo and gradient descent
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Hardness of Learning Fixed Parities with Neural Networks
Any fixed parity of size at least logarithmic in the dimension requires exponentially many perturbed-gradient steps before the expected correlation loss moves away from its trivial value.